Insight ON From Assistant to Autonomy: Navigating the Three Levels of AI in Software Engineering
By Insight UK / 24 Nov 2025 / Topics: Artificial Intelligence (AI) Software
Executive summary
If you are responsible for custom built business systems, your world is turning upside down. AI is revolutionising the business of software engineering. The practices you’ve been using for 25 years must now be completely re-designed.
Software Engineering Services
Some custom software is built in-house, but $200 billion per year is spent contracting out software engineering to consultancies and outsourcers. Very few contracts are based on results, though. Most are priced based on hours consumed.
AI and Software Engineering
Since 2020 there has been an explosion of tools which use generative AI for software engineering. OpenAI, Google and Microsoft have all produced tools which can generate code; and there is a host of fast-growing startups with competing offerings.
Business Impact
AI can save 40%-50% of software engineering effort. (Though reports vary - in some situations the benefits are much smaller).
That means there is a huge opportunity to deliver business change faster. Requests from the business should no longer have to wait in a backlog for months. Teams can deliver working software with new substantial features in days.
To get there, however, we must change the way the business works with software engineers. We must make software engineering teams smaller and re-think the roles in an agile team.
